Pontocerebellar hypoplasia (PCH) is a heterogeneous group of diseases characterized by lack of development and/or early neurodegeneration of cerebellum and brainstem. Indeed, PCH has been associated with chromosomal abnormalities or metabolic disorders as well as congenital muscular dystrophies 1 2 3 4 5 . According to clinical features, seven subtypes of PCH have been described (Table 1) 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 .

Recently, the identification of several novel responsible genes has added new insight to the characterization of this group of disorders: three out of four subunits of the tRNA splicing endonuclease complex, TSEN54, TSEN34 and TSEN2 have been shown to be involved in PCH2, PCH4, PCH5 and in one patient with PCH1 10 12 17 . PCH6 as well as some PCH1 are related to mutations in the mitochondrial arginyl-tRNA synthetase RARS2 gene 18 19 20 . The VRK1 gene was identified as responsible for PCH1 but the reported patients seemed to be affected by a very mild condition compared to the clinical description of classical PCH1 10 . In addition, similar neuroimaging aspects of the cerebellum and brainstem have been observed in patients with large deletions/mutations of the VLDLR gene and in Reelin-associated lissencephaly with PCH 22 23 24 25 .

More recently de novo anomalies in the X-linked gene CASK (calcium/calmodulin-dependent serine protein kinase) have been identified in patients with PCH, mostly females, displaying intellectual disability and microcephaly (MICPCH syndrome) 26 27 28 29 .

In this study, we performed a systematic screening for CASK mutations and deletions in 14 patients presenting with PCH at neuroimaging but for whom the clinical features differed from PCH1 or PCH2 (PCH4 and PCH5). We observed a high frequency of CASK abnormalities (13/14). We describe the clinical aspects of these patients, including the second reported male patient with an extremely severe PCH and clinical phenotype and the first mosaic mutation found in a mildly affected male. Finally, we discuss the variability of the clinical and radiological phenotypes and highlight the characteristics of this recognizable syndrome.

Overall, a CASK anomaly was identified in 13/14 patients (10 intragenic mutations and three Xp11.4 submicroscopic deletions).

Calcium/calmodulin-dependent serine protein kinase (CASK) is a member of the membrane-associated guanylate kinase (MAGUK) family involved in synapse formation and in the regulation of gene expression, including Reelin which is critical in brain development 30 .

Using Array-CGH, two groups identified Xp11.4 submicroscopic deletions involving the CASK gene in girls affected with intellectual disability and microcephaly 31 32 . Najm et al. reported three novel female patients with CASK deletions or disruption and described more precisely the phenotype including cerebellar and brainstem hypoplasia 26 . Furthermore in two patients without Xp11.4 deletions they identified de novo CASK intragenic mutations. One of the patients was a boy who was affected with a very severe form of PCH. Thirty-three further female patients with PCH and CASK de novo loss-of-function anomalies have been reported 27 28 29 . In addition, CASK hypomorphic inherited mutations have been described in patients, mainly male, affected with intellectual disability and not featuring MICPCH 33 34 35 .

In this study, we describe 13 patients (11 girls and 2 boys) affected with PCH and harboring heterozygous or hemizygous CASK anomalies. Among a large series of PCH patients that were referred to us, we prioritized the screening of the CASK gene in these patients since clinical criteria for PHC1 and PCH2 were absent. We particularly noted the differences in the clinical features between our patients and those of PCH2 patients. Indeed, patients had no incessant choreic movements, were hypotonic rather than hypertonic, had a less severe psychomotor retardation and had features not usually observed in PCH2 (sensorial involvement as well as slight dysmorphism). Sequencing allowed us to identify 10 CASK intragenic mutations, in eight girls and two boys. All mutations were private, with no identified hotspot. Nine were inactivating mutations and one was predicted to lead to a large in-frame deletion. In three other female patients Array-CGH analysis revealed a Xp11.4 submicroscopic deletion involving part of or the entire CASK gene.

The patients with a CASK mutation showed a characteristic phenotype with some variability concerning the severity and the associated sensorial involvement. All had neonatal feeding difficulties, intellectual and motor disabilities and progressive microcephaly. Most had profound disabilities but two patients are able to use a few words and have acquired autonomous walking. Postnatal microcephaly was severe in seven patients (OFC < -6SD) but OFC ranged between -3 to -4 SD in four. Four patients had epilepsy and this was refractory in one. Ophthalmological abnormalities were present in at least six patients and sensorineural hearing loss in two. The association with a diabetes mellitus observed in one patient has not been previously reported and could be co-incidental. All our patients with a CASK mutation presented slight but homogeneous dysmorphic features. As reported by Moog et al., we observed a broad nasal bridge in the majority of patients and a micrognathism in the younger. A round face tending to lengthen in older patients, well-drawn arched eyebrows and protruding maxilla and maxillar incisors in older patients represented the major facial findings in our series. These findings were also frequently observed in patients presented by Najm et al and Moog et al 26 29 .

In one patient we failed to detect any CASK mutation or deletion. This patient had a very severe phenotype characterized by severe encephalopathy, almost no psychomotor acquisition, epilepsy, absence of choreoathetosic movements and premature death. MRI revealed severe pontocerebellar hypoplasia as well as severe subtentorial atrophy. A previous pregnancy has been terminated because of PCH suggesting the possibility of recessive disease.

All our patients had severe intellectual and motor disabilities with PCH but at varying degrees. A correlation between the clinical severity and the MRI findings may be discussed. This correlation is obvious for the two male patients, notably due to the severity of patient 13. He had the most severe clinical features including refractory epilepsy, optic atrophy and almost no psychomotor acquisition at age of 16 months and his MRI showed the most severe PCH. By contrast, the male with a mosaic mutation (patient 12) was able to walk from the age of three years and use a few words. He had no epilepsy or spasticity, his microcephaly was moderate (-3.5 SD) and his MRI showed mild cerebellar hypoplasia. This correlation is less clear however in female patients. Nevertheless, the female patient with the mildest clinical features, patient 10, had a clinical phenotype very similar to that of the male patient 12 and also had mild cerebellar hypoplasia. Moreover, in previously reported female patient series the acquisition of walking and the ability to use a few words seems to correlate with a less severe cerebellar hypoplasia (Patient 6 in Hayashi at al.; patients 7, 15, 20, 21 and 24 in Moog et al) 28 29 . However this clinicoradiologic correlation cannot be relied upon entirely since patient 11, while still young, had a mild cerebellar hypoplasia but she was unable to walk. As reported by Takanashi et al., we observed a reduced area of the cerebrum, pons and cerebellum, but a sparing of the corpus callosum in most patients 36 .

Aside from this partial clinicoradiologic correlation, we failed to identify any clear genotype-phenotype correlation in this series, except for the non mosaic male patient, who is hemizygous for the splice mutation and has the most severe phenotype. The severity of the disease in patients with a CASK deletion was very similar to patients with intragenic mutations regarding motor and intellectual disability, epilepsy, MRI anomalies and hearing and visual defects, even when the deletion involved nine other genes. Furthermore, two very close mutations (p.Trp657* and p.Trp656*) were observed in two unrelated girls: one had the milder phenotype, acquired independent walking at age 3, was able to run and to use a few words at 8 years and the second was a severely affected girl who had acquired no language and was unable to walk.

Finally, all girls and boys with inactivating CASK mutations presented with a recognizable phenotype characterized by severe postnatal microcephaly, moderate to severe motor and intellectual disability, feeding and sleep difficulties, frequent ophthalmological anomalies and PCH on MRI. The absence of anterior horn involvement excluded PCH1. Severe post-natal microcephaly and dystonia as well as some neuroradiological features (dragonfly-like cerebellum, hypoplasia of the pons) could be suggestive of PCH2. However, PCH2 and CASK-related PCH are distinct entities. Indeed, PCH2 patients have a more severe clinical course (near absent development and choreic movements), and do not present with ophthalmologic abnormalities (glaucoma, megalocornea, optic atrophy, and retinopathy), hearing loss or facial dysmorphism, which are frequently found in CASK mutated patients. Furthermore, analysis of the MRI gives additional clues showing relative sparing of the pons, asymmetry of the cerebellar hemispheres as well as vermian and cerebellar involvement. In PCH2 there is a relative sparing of the vermis but cerebellar hemispheres and brainstem are severely affected with a flat pons 9 29 37 . If there is any doubt, the major TSEN54 mutation (A307S) observed in a majority of PCH2 patients, can easily be ruled out.

Only three male patients with MICPCH and de novo CASK mutations have been reported [this report, 26] suggesting a possible lethality in most affected male patients. The two male patients with MICPCH and non mosaic CASK mutations presented with a severe PCH, severe neurological impairment with near absent development, absence of sucking and swallowing, optic atrophy, refractory epilepsy and edema of the hands and feet in one 26 . This phenotype is reminiscent of PEHO syndrome but the presence of severe PCH unlike progressive cerebellar atrophy should prompt to look for CASK mutations 38 .

To date, two groups have reported molecular analysis results in their cohort of PCH patients 9 27 . These studies showed that the majority (45 to 63%) of patients were affected with PCH2/4 with mutations in one TSEN gene, mostly the A307S mutation in TSEN54. In our cohort of 40 PCH patients referred to our reference center, CASK-related PCH was the second most frequent cause (13/40, 32%) behind PCH2/4 (18/40, 45%, manuscript under preparation).

Several groups have reported CASK mutations in male patients with other phenotypes: FG syndrome and X-linked mental retardation (XLMR) with or without nystagmus 33 34 35 . The mutations reported in these patients were missense or splicing mutations resulting in in-frame deletions and were all familial. The affected boys had mild to severe intellectual disability without severe motor impairment and without microcephaly. Nystagmus was a frequent feature and some patients had seizures, strabismus, optic disc pallor, visual impairment or dysmorphic features. In those families, most females had no (10/13) or mild (2/13) intellectual disability and one was severely retarded. Data concerning MRI are available in only 2/25 male patients, one of whom has been described as normal and the other showing cerebellar hypoplasia and pachygyria 35 . The male patient with cerebellar hypoplasia had severe intellectual disabilities, nystagmus and a small OFC at the 3^rd centile, but no motor abnormalities or seizures were reported and his mutation was inherited. More details concerning the MRI features of male patients with hypomorphic CASK mutations would be contributive to determine which mentally retarded patients should be tested for CASK mutations.

Therefore the spectrum of phenotypes associated with CASK mutations is very large in males, from lethality or severe neurologic impairment with PCH to isolated intellectual disability. In females, depending on the type of mutation, it extends from normal to severely mentally retarded patients with motor disabilities related to PCH.

Genotype-phenotype correlations in our patients and previously published patients confirm that CASK inactivating mutations are de novo and are associated with PCH and microcephaly in females and a severe phenotype in males while missense mutations or in-frame deletions are observed in males and rarely in females with variable degrees of XLMR with or without nystagmus.

As patients with non-inactivating mutations have a non specific intellectual disability, it seems difficult to determine the clinical criteria which would allow the selection of patients for sequencing, particularly in sporadic cases. On the other hand, patients with inactivating mutations have a recognizable phenotype consisting of a specific form of PCH. Consequently CASK analysis should be performed in girls presenting with intellectual disability, microcephaly and PCH, in the absence of symptoms suggesting PCH2 (severe encephalopathy with choreoathetosis) and particularly in the presence of ophthalmological anomalies such as optic atrophy or glaucoma. Furthermore we consider that CASK exonic deletion or duplication screening must be part of the strategy when looking at CASK anomalies in a patient with a suggestive phenotype. The identification of a de novo mutation allows to give a reassuring genetic counseling, whilst PCH is most often autosomal recessive.

PCH: Pontocerebellar hypoplasia.

DL is on scientific advisory board of Genzyme. The other authors declare that they have no competing interests.

LB designed and coordinated the study, saw some of the patients, revised all the clinical and MRI data, carried out the molecular genetic studies and wrote the manuscript. SCB and AC carried out Array-CGH analysis and co-wrote the manuscript. CGa revised all MRIs and co-wrote the manuscript. RT performed X-inactivation study, collected the clinical data of one patient and revised the manuscript. MM, FP, AA, CGo, SB, CI, LL, SJ, ID, DL, VDP contributed to the clinical work, seeing patients and collecting clinical data. GZ and EB contributed to the acquisition and analysis of the clinical and molecular data in one patient, and co-wrote the manuscript. JPS and TBV contributed towards analyzing the data and revising the manuscript. DR designed and co-coordinated the study, contributed to the acquisition and analysis of the clinical and MRI data and co-wrote the manuscript. All authors read and approved the final manuscript.